feat: add vLLM serving, streaming, and Ray Serve configuration - #12
Merged
Merged
Conversation
Dispatches to a real vLLM server, serving a selected LoRA adapter by name over the shared base model, parses usage and finish reasons, streams content deltas, and reports engine health. Unreachable engines and unusable bodies raise an explicit backend error.
Emits OpenAI-compatible SSE chunks, caches and replays streamed results under the same eligibility rules, fails an unreachable engine with 502 and retry guidance, and makes readiness depend on backend health.
Renders per-tier autoscaling, GPU pool placement, tensor parallelism, prefix caching, quantization, and Multi-LoRA settings from model cards, plus a canary description naming its rollback target. The committed config/ray-serve.yaml is verified against the catalog by a test.
github-actions
Bot
added
documentation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Implements sections 7.3, 7.4, and the streaming half of 7.1.
VLLMBackenddispatches to a real vLLM OpenAI-compatible server and serves a selected LoRA adapter by name; unreachable engines return 502 with retry guidance and fail readiness.config/ray-serve.yamlis generated from the catalog (autoscaling, GPU pools, tensor parallelism, prefix caching, quantization, Multi-LoRA) and a test fails if it drifts.Tests: 19 unit, 6 integration, 1 Playwright end-to-end. Local suite 88 passed.